PROTECTIVEEFFECT OF VACCINATION AGAINSTINDUCED INFLUENZA A'

BY THOMASFRANCIS, JR., JONAS E. SALK,2 HAROLDE. PEARSON,ANDPHILIP N. BROWN

(From the Department of Epidemiology and Virus Laboratory, School of Public Health,University of Michigan, Ann Arbor, and the Ypsilanti State Hospital,

Ypsilanti, Michigan)

(Received for publication December 19, 1944)

In December 1942, rather extensive studies,which will be reported separately, were begun forthe purpose of determining the value of subcu-taneous vaccination with a vaccine of inactivatedinfluenza viruses, Types A and B, against theepidemic form of influenza. The anticipated out-break of this disease did not occur but the oppor-tunity to test the resistance of certain of the vac-cinated individuals to induced infection presenteditself. This was done by inducing infection withstrains of virus maintained under laboratory con-ditions. Other investigators (1) had previouslyreported that, by the use of such a procedure, vac-cine containing inactivated Type A influenzavirus, given subcutaneously, was demonstrated tohave protected children against experimentalinfection.

A preliminary report of the clinical results hasalready been made (2). The present report isintended to present the details of the evidence forthe protective effect of vaccination against inducedinfluenza A, including the laboratory investiga-tions which constituted an important phase of thestudy.

MATERIALS AND METHODS

Vaccine. The vaccine employed was prepared by themedical research division of Sharp & Dohmes accordingto specifications furnished by the Influenza Commissionand purchased at a minimal cost with Commission funds.

1These investigations were aided through the Com-mission on Influenza, Board for the Investigation andControl of Influenza and other Epidemic Diseases in theArmy, Preventive Medicine Service, Office of the Sur-geon General, United States Army. This study was alsoaided by a grant from the International Health Divisionof the Rockefeller Foundation.

2Fellow in the Medical Sciences of the National Re-search Council 1942-1943.

8 The authors wish to express to Dr. Wm. A. Feirerand Dr. Bettylee Hampil an appreciation of their activeinterest and cooperation in the preparation of the material.

As a source of virus, 11-day-old embryonated eggs wereinoculated with a 10-3 dilution of infectious allantoic fluid.The PR8 strain (3) of Type A virus and the Lee strain(4) of Type B virus were used. The virus contained inthe harvested extra-embryonic fluids was adsorbed ontothe embryonic erythrocytes and concentrated by elution in

%lo the original volume of saline (5). The Type A andType B materials were prepared separately and tested forvirus content by agglutination of chicken red cells andalso by infectivity tests in mice. The virus was inacti-vated by the addition of formalin in a final concentrationof 1: 2000 and phenyl mercuric nitrate, 1: 100,000, wasadded for bacteriostatic purposes. The Types A and Bviruses were pooled in equal volumes and after bulksterility tests were bottled in 50 ml. volumes. The otherstandard tests as specified by the National Institute ofHealth for sterility and safety required for biologicalproducts were then made.

The immunizing capacity (10) of the vaccine for micewas such that 2 doses of 0.5 ml. each of a 2 X 10-2 dilu-tion of vaccine, given intraperitoneally a week apart,protected mice against at least 10,000 50 per cent mor-tality doses of mouse passage PR8 virus, given intra-nasally, 1 week after the last intraperitoneal injection.

Virus preparation used for infection. The virus usedfor inducing infection in the human subjects was theBaum strain of Type A influenza virus isolated in 1940in ferrets inoculated with throat washings from a patientacutely ill in the epidemic. The strain is similar to butnot antigenically identical with the PR8 strain of TypeA virus present in the vaccine. Frozen and dried mouse

lung tissue, which had been stored for nearly 2 years,served as seed for preparation of the infected allantoicfluid employed for the human infection. The virus hadbeen passed through 9 ferrets and 10 series of mice beforeegg inoculation. Material from the 3rd egg passage wasused. Fluid without red blood cells was aspirated fromthe allantonic sac of 13-day-old embryos 48 hours afterinoculation. The material was clarified by centrifugation,placed in rubber capped vaccine vials, and kept at 40 C.for 20 days before use. Tests for bacterial sterility weremade in Brewer's thioglycollate medium, in plain beefheart infusion broth, and by inoculation of 11-day-oldchick embryos. At the time the fluid was used for hu-man infection, it had the following properties:

(1) Intranasal infective titer for mice:

536

VACCINATION AGAINST INFLUENZA A

10 103 10-' 10' 10- 10-73,4,4 3,5,5 5,6,6 5,6,7 +++,++,= +,+,+

(Figures denote day of death of individual mice, and plussigns refer to degree of pulmonary consolidation in survivorsautopsied on the 10th day.)

(2) Agglutinating titer for chicken's red cells was 1280(6, 7).

Subjects. The subjects were 102 male residents of asingle ward of the Ypsilanti State Hospital, Ypsilanti,Michigan. The men ranged in age from 20 to 70 years,but the majority were between the ages of 30 and 54years. For the most part, they were able to give ade-quate descriptions of their symptoms. They were physi-cally active and almost all were productively employed inthe various maintenance services of the institution. Whennot at their jobs, the men mingled in a large, modern,well ventilated day room and at night were quartered inspacious dormitories containing 20 to 30 beds each withspace of approximately 3 feet between beds; a few menhad individual rooms. They took their meals in a cafe-teria together with men from other wards.

Vaccination. On December 21, 1942, alternate resi-dents of the ward received a subcutaneous injection of 1ml. of the combined Types A and B vaccine. In thesame manner and at the same time, the others were given1 ml. of physiological salt solution containing formalinand preservative in quantities similar to that present inthe vaccine. The subjects did not know which materialthey had received. Careful daily observations were madeduring the subsequent 4% months during which time noevidence of influenza A was found.

On April 21, 1943, the ward population comprised 45individuals who had received vaccine in December 1942,and 57 who were either uninoculated or were given aninoculation of control material. At this time, 4 monthsafter the initial vaccination, 17 of the 45 subjects in thevaccinated group were given a second inoculation of thesame lot of vaccine which had been administered in De-cember. At the same time, 21 of the 57 in the unvacci-nated group received an initial injection of vaccine. Thevaccine had been stored at 40 C. since preparation.

Infection. On May 4, 1943, all subjects were exposedin a uniform manner to influenza virus Type A. Dividedaccording to prior treatment, the ward population con-sisted of the following groups:

(1) A control group of 36 unvaccinated individuals;(2) 28 men who had received vaccine 4% months

previously;(3) 17 men who were twice vaccinated, both 42

months and 2 weeks prior to infection;(4) 21 individuals who had received vaccine 2 weeks

before infection.All subjects were kept from their usual occupations

for 24 hours following exposure to the virus; however,as a group they visited the cafeteria for their meals. Pa-tients from other wards used the same dining hall at thesame time, but intermingling of groups was avoided.After the first day, all except those who were febrile wereallowed to continue regular activity and to take meals in

the cafeteria. Although quarantine precautions were notrigid, no evidence of influenza was observed at this timein any other group within the institution.

The method for introducing the virus, using nebulizers 4for dispersing the virus as a very fine spray, was the sameas that previously described (8). The spray was directedby means of an adapter into each nostril for 2 minutesand the subject allowed to breathe the mist. Under apressure of 10 pounds of air, the nebulizers deliveredapproximately 0.5 ml. of fluid during the 4 minutes' expo-sure. The procedure was carried out on the ward in atreatment room directly off the corridor leading to thedormitories and day room.

Since there was a free interchange of air between allrooms, virus which escaped into the atmosphere duringthe course of the spraying was widely disseminated. Thiswas demonstrated by the isolation of virus from the lungsof mice exposed at several locations in the quarters.While the inhalation procedure was in progress, groupsof 10 mice in wire-mesh cages were placed in the treat-ment room as well as in two other locations on the ward,about 25 to 30 feet away. Four days after exposure, 5mice in each group were given an intranasal inoculationof sterile broth to accentuate the effect of virus that mightbe present in their lungs (9). The fate of these micewas as follows:

Location

Treatment roomCorridorDay room

Exposed during spraying6, 6, 6, 7, 9x ++, +, +,4-+,0, 0,0, 0

Exposed duringspraying and in-oculated intra-

nasally withsterile broth4 days later

6,6,6,67,77, 7, 7, 9, 97,9,0,0,0

(Numerals denote day of death after exposure. Symbols0 to + +++ indicate degree of pulmonary involvementin survivors autopsied on 10th day.)

x only 4 mice in this group.

It is seen that virus was present in sufficient concentra-tion in the atmosphere in different locations of the wardto infect mice exposed for several hours.

Twenty-four hours after the spraying, 10 mice wereexposed in the treatment room for 30 minutes, and thentreated as above. No pulmonary lesions were observed.

Clinical observations. All subjects were examined forevidence of respiratory disease on the day before exposureto the experimental infection. Sublingual temperatureswere taken for 5 minutes at that time and again immedi-ately before inhalation of virus. Subsequently, all- wereobserved twice daily by physicians and nurses. Tempera-tures, symptoms, and signs of illness were recorded. Ob-servers did not know to which group the individual sub-jects. belonged.

White blood cell counts were obtained from 21. of thesubjects on the day before infection and then daily for4 days.

Roentgenograms of the chests of 3 individuals who had

4Nebulizers for use in administering epinephrin weremade by G. E. Miller, Philadelphia, Pennsylvania.

537

T. FRANCIS, JR., J. E. SALK, H. E. PEARSON, AND P. N. BROWN

the most pronounced symptoms and fever were taken onthe 3rd day of illness.

Serology. Blood for serological study was obtainedfrom all individuals vaccinated on April 21, 1943. OnMay 3, 1943, the day before exposure to virus, samplesof blood were taken from all subjects for determinationof the titer of antibodies in the serum before infectionand for estimation of the response to vaccine in the inocu-lated individuals. Blood was again drawn 2 weeks afterinhalation of the active virus.

Antibody was titrated by measuring the capacity ofserum to inhibit the agglutination of chicken's erythro-cytes by influenza virus. The technique followed is amodification (7) of Hirst's (6) procedure. The serawere titrated against the PR8 strain of influenza virus,Type A. A single pool of antigen was used throughout.Some sera were also tested with the Baum strain whichhad been used for infection. No important differences inserological responses were noted.

CLINICAL PICTURE OF EXPERIMENTALLY

INDUCED INFLUENZA A

The clinical picture of the experimentally in-duced infection resembled a mild form of the natu-rally occurring disease. The most common initialsymptoms were chilliness, headache, generalizedbody aches, weakness, and cough associated withsubsternal soreness. A few individuals com-plained of slight irritation of the throat, mildnasal congestion, and aching eyes. Dizziness andnausea were each recorded once. Anorexia andinsomnia were common. All degrees of severityof illness were observed, varying from the com-plete syndrome, with a sharp febrile reaction, tomild indisposition associated with definite butslight achiness and headache with little or nofever.

While temperatures of 990 to 99.80 may havereflected a mild illness, these weres difficult toevaluate since they were noted irregularly in indi-viduals before infection. Correlation betweensymptoms and degree of fever revealed that tem-peratures of 1000 F. or more could be consideredindicative of distinct clinical disease and a tem-perature of less than 1000 was therefore not con-sidered to be evidence of infection.

In the group of unvaccinated subjects, thehighest temperature observed was 1030 F. Inthe majority of those who developed distinct fe-brile reactions, temperatures of 1000 or more werefirst noted within 24 hours after inhalation; inthe remainder, fever began between 24 and 48hours after exposure to the virus. On the basis of

temperatures taken twice daily, elevations of 1000or more lasted in most instances no longer than24 hours after onset, in a few persisted for 48hours, and for as long as 60 hours in only onesubject. In no instance in which the maximumtemperature recorded was less than 1010 did feverlast longer than 24 hours. However, the ma-jority of those with temperatures of 1010 or morehad fever persisting more than 24 hours. Nodifference was apparent in the features of illnessobserved in vaccinated and control subjects whencases with corresponding degrees of fever werecompared.

Physical examination revealed flushed skin andvarying degrees of prostration, usually parallelingthe height of fever. Spasmodic, dry cough com-monly appeared early. In some, on the secondday of illness, slight redness of the pharyngealand nasal mucosae was evident, occasionally ac-companied by nasal congestion. Abnormalphysical signs were detected in the chests of 3of the subjects with the most marked reactions.The findings, over localized areas and involvinga major portion of a lobe, were medium, high-pitched, sibilant rales, more marked during theexpiratory phase of respiration, together withscattered, fine, moist rales, audible during bothinspiration and expiration. Breath sounds werediminished slightly in intensity and were ofbronchovesicular quality. No obvious alterationin percussion note accompanied the auscultatoryfindings. Roentgenograms of the chest on the3rd day of illness revealed no abnormalities inany of the 3. The abnormal physical signs dis-appeared within 1 week.

In the most severe cases, a moderate degreeof asthenia was present for several days aftersubsidence of fever, and a slight spasmodic, non-productive cough persisted in many for as long asa week. On the whole, the disease was mild withprompt and uneventful recovery.

White blood cell counts taken on days follow-ing inoculation tended to be lower in comparisonwith the pre-inoculation level. This trend wasnot uniform nor did it occur to a marked degree.

DISTRIBUTION OF CLINICAL DISEASE IN CONTROLAND VACCINATEDGROUPS

Analysis of the clinical responses of the dif-ferent groups to inhalation of virus revealed dis-

S38

VACCINATION AGAINST INFLUENZA A

TABLE I

Effect of subcutaneous vaccination upon febrile response of human subjects to induced infection with influenza virus, Type A

Highest temperature

NumberVaccination record of <99 99 + 100 + 101 + 102 +

subjects

No. Percent No. Per cent No. Percent No. Per cent No. Percent

Unvaccinated 36 7 19 29 81 18 50 9 25 4 114} mos. before 28 8 29 20 71 9 32 3 11 1 42 weeks before 21 7 33 14 67 3 14 0 0 0 041 mos. and 2 weeks before 17 6 35 11 65 3 18 0 0 0 0

tinct and significant differences (2). The resultshave been expressed in terms of the number ofsubjects reacting with temperatures above thevarious levels in order to emphasize the observa-tion that among the several groups there is adifference not only in the incidence of tempera-tures above 1000 F., but in the severity of illnessas well, as is indicated by the height of the febrileresponse. Table I summarizes the maximum tem-peratures recorded. Temperatures of 1000 ormore were observed in 50 per cent of the con-trols, in 32 per cent of those vaccinated 41/2months before exposure, in 14 per cent of thosevaccinated 2 weeks before, and in 18 per cent ofthose vaccinated both 41/2 months and 2 weeksbefore; the incidence in the 2 groups, totalling 38individuals, who had been vaccinated 2 weeksbefore infection was 16 per cent. Moreover, noneof the latter had temperatures higher than 100.80,while 25 per cent of the control group and 11per cent of the men vaccinated 4Y2 months earlierhad fevers of 1010 or higher. Febrile reactionsof 1020 or more were observed in 11 per cent ofthe controls as compared with 4 per cent of thegroup vaccinated 41/2 months before infection.

The immunizing effect seemed to be moremarked in the 2 groups vaccinated within 2 weeksof exposure. Individuals receiving 2 inoculations4Y2 months and 2 weeks prior to exposure ap-peared to have had no advantage over those re-ceiving a single inoculation within the shorterinterval. The group vaccinated but once, 4Y2months before infection, was not as resistant aseither of the other 2 vaccinated groups, suggestingthat immunity may have waned in the 4½2-monthinterval between vaccination and exposure to in-fection. It is interesting to note that the distinctclinical reactions in the former group occurred

chiefly among those who had the lower titers ofserum antibody.

SEROLOGICALRESPONSESOF VACCINATEDAND

REVACCINATEDSUBJECTS

Antibody responses were determined in sub-jects receiving their initial injection of vaccine inApril 1943 as well as in those subjects who werereinoculated at that time, 4 months after their firstinjection. Antibody response to the vaccinationdone in December 1942 was studied in only 10per cent of the population vaccinated at that time.Consequently, the number of subjects in the pres-ent group who were tested following vaccinationare too few to discuss. Since antibody responseof human subjects to vaccination varies widely,studies of the antigenic activity of vaccines mustbe done in large groups of individuals. For thisreason, the results obtained in the small groupsinvolved in this experiment can be considered toindicate, only in a general way, the antigenic ef-fect of the vaccine. Studies on larger groups,spread throughout a greater population will bereported subsequently (11).

As shown in Figure 1, 17 of the 21, or 81 percent, of the group vaccinated for the first time 2weeks before infection had 2-fold or greater in-crease in antibody titers and 12, or 57 per cent,had increases in titers of 4-fold or more. In therevaccinated group, 6 of 16 tested, or 38 per cent,had 2-fold or greater increases in titers, and only2, or 13 per cent, had greater than 2-fold in-creases. Both of the latter had titers of less than32 prior to the second injection. As has beenobserved previously, changes in titers occur lessregularly and are of relatively slight degree inindividuals who have high antibody titers at thetime of injection (12). Revaccination appeared

539

T. FRANCIS, JR., J. E. SALK, H. E. PEARSON, AND P. N. BROWN

409 0*

z20 * *0

z- 512 _

hi 01 25 0 0 * .0

128

_ 6X *I0-

430 *0U

I I I . I f1 <3 <32 32 64 128 256 512

ANTIBODY TITER BEFORE INJECTION

A

4094

204

102

SI

25

12

6

0

2 0@0

* .0

0

I I I I I I<32 32 64 128 256 512 1024

ANiIBODY TITER BEFORE INJECTION

BFIG. 1. A. CHANGEIN TYPE A ANTIBODY TITER FOLLOWINGFIRST INJECTION OF

VACCINE. B. CHANGEIN TYPE A ANTIBODY TITER FOLLOWING SECONDINJECTIONOF VACCINE, 4 MONTHSAFTER FIRST

to have had only a slight influence on the anti-body levels of the group as a whole. In this re-spect, reinoculation of a group of adults, whoprobably have had multiple experiences with in-fluenza viruses, did not seem to have a "booster"effect.

RELATION BETWEENTITER OF SERUMANTIBODYAND RESISTANCE TO INFECTION

The data obtained in this experiment were ex-amined to determine whether or not resistance toinfection, as measured by febrile reaction, was, toany degree, related to the -level of serum antibody.The relationships are shown in Figures 2 and 3.

In general, the data in Figure 2 reveal that inthe unvaccinated group antibody titers were inthe lower range, while in the vaccinated groupsantibody titers were higher. Moreover, the inci-dence and degree of distinct febrile reactions weregreater in unvaccinated subjects as comparedwith vaccinated individuals. It is readily evidentfrom these results that no striking relationshipbetween antibody level and resistance is apparentin any one of the 4 groups, and that if any rela-tionship does exist, the number of individuals ineach group is too small to show it conclusively.It can be pointed out that in the group vaccinated4Y2 months before infection, 6 of 9 who developed

distinct febrile responses had antibody titers of128 or less, while 3 out of 17 who had titers of256 or greater had distinct fever. On the otherhand, of the group vaccinated 2 weeks before in-fection, there were 8 out of 9 with titers of 128 orless who failed to exhibit a clinical response, while10 out of 12 with titers of 256 or greater did notrespond with a significant degree of fever.

When a composite chart is made, by super-imposing all 4 groups, it is seen that those whohad febrile reactions tended to have the lowerantibody titers. Among those with titers of 256or more, there were few who had a febrile re-sponse. When the incidence of temperatures of1000 or more is determined at each level of serumantibody before infection, distinct differences be-come evident. Table II summarizes the observa-tions illustrated in Figure 3. Temperatures of1000 or more were noted in 26, or 49 per cent,of 53 individuals having pre-infection antibodytiters of 128 or less; in 7, or 19 per cent, of 36with titers of 256 and 512; and in none of the 13having titers of 1024 or greater. When a sepa-ration is made between those with titers of 128or less, and 256 or more, dividing the group ap-proximately in half, temperatures of 1000 or morewere found to have occurred in 49 per cent ofsubjects in the low antibody group as compared

540

VACCINATION AGAINST INFLUENZA A

0 UNVACCINATED CONTROLS

00

00

00

* 0

00 0_~~~~~~~I_ 4_______

*-00

00 0 0

0 0

00

00

I0

11030a:

h 102

I.-<lot

hi

2I.-

2

2

2 8<32 32 64 28 256 512 002420484096ANTIBODY TITER BEFORE INFECTION

0 VACCINATED 4 MOS.BEFORE INFECTION

0

0

000*

*00* ~~0 0 0* * 0

* 0** 0

I I I

00 0 0 00* *

1

<32 32 64 .128 256 512 1024204 4095ANTIBODY TITER BEFORE INFECTION

VACCINATED 2 WKS.BEFORE INFECTION

0 00

" 1034:

16 102

t 101

204c

22 942K,4c2 94

0 *0* 0 00 0

0 0

(32 32 64 128 256 512 1024 20484096ANTIBODY TITER BEFORE INFECTION

VACCINATED TWICE4% MOS. AND 2 WKS.BE ORE INFECTION

0 0 0

*-

* 00

*000*

.0

<32 32 64 128 256 512 10242048409ANTIBODY TITER BEFORE INFECTION

FIG. 2. RELATION OF ANTIBODY TITER AND FEBRiLE RESPONSETO INFLUENZA VIRus, TYPE A

101 *o

0

0

0

@ UNVACCINATEDCONTROLSO VACCINATED 4¶, MOSBEFOREINFECTIONO VACCINATED2 WKS. BEFOREINFECTIONx VACCINATEDTWICE4J MOS. AND2 WKS

BEFORE INFE CTION

*. 00

0*x x 6

0

oO- - -0*-O- - - --0____ __

O*0 0 0* 0

o 0*0 OXXX* 00

00 0*000 00

x xoao0

0

X 0

00 xxOO 00xx.xx- 0

00 0

0 00

0 X

<32 32 64 128 256 512 KANTIIBODY TITER BEFORE INFECTION

024 2048 4096

FIG. 3. RELATION OF ANTIBODY TITm AND FEBRILE RESPONSETO INFLUENZA VIRus, TYPE A,IN VACCINATEDAND CONTROLSUBJECTS

0 0

0 103

101102

x

A-g.. 100

2

2

2

541

r 103A.

W 102I-

Z 4

er 101

2

2, 992

2 98

0

.0----0-- - - - --0 0

0

* 0

0

00K

^, 102

< 0

A.

r

hi

-10

.4a:

x

A.2us1--

100

4I

0 x 00*0-

I I~~

99C

rt

T. FRANCI$, JR., J. E. SALKE, H. E. PEARSON, AND P. N. BROWN

TABLE II

Ixcidence of febrile reactions of 100° or more in relation topre-infection antibody titer

Incidence of temperature of 1000 or more

Antibody titerNo.* Per No.* PrNo.* Pe

cent cent cent

32 or less 4/8 50

64 12/25 48 26/53 49 26/53 49

128 10/20 50

256 4/21 197/36 19

512 3/15 20 7/49 14

1024 or more 0/13 0 0/13 0

* Numerator = Number of individuals having tempera-ture of 100° or more.

Denominator = Total number in each group.

with 14 per cent of individuals having the higherlevels of antibody. Only one of the latter had asmuch as 1010, while 11 of the former had 1010or more.

Because of the preponderance of unvaccinatedsubjects in the low antibody group and of vacci-nated individuals in the higher antibody group,as well as the fact that the vaccinated subjects donot represent a uniform group with regard tovaccination experience, the data do not permit aconclusive statement regarding the relationshipbetween antibody level and resistance.

SEROLOGICALDIAGNOSIS OF INFECTION IN

VACCINATED SUBJECTS

In the course of these studies with experi-mentally induced infections, cases of clinical in-fluenza have been observed both with and withoutcorroborative serological reactions. When thenumber of cases showing dissociation betweenclinical reaction, as measured by febrile response,and serological change, are compared in the con-trol and vaccinated groups, distinct differencesare evident. Figure 4 relates maximum tempera-ture to antibody response following the inhala-tion of virus. These results are summarized inTable. III.

These data clearly reveal a greater incidenceof positive serological reactions following the in-halation of active influenza virus in unvaccinatedas compared with previously vaccinated subjects;

TABLE III

Relation of serological response to febrile reaction followinginhalation of active virus in control and vaccinated subjects

Temperature reaction

1000 + < 1000Antibodyresponse Controls Vaccinated Controls Vaccinated

(18) (15) (18) (61)

No. Per No. Per No. Per No. Percent cent cent cent

None 2 11 8 53 10 56 48 79Two-fold or 16 89 7 47 8 44 13 21

greaterFour-fold or 13 72 1 7 4 22 2 3

greater

a similar relationship was found when a divisionwas made on the basis of low and high antibodylevel before infection. This is true both amongindividuals who exhibited little or no clinical re-action, as measured by temperature, as well asamong those showing a distinct clinical response.The fact that antibody rises can occur in theabsence of any clinical evidence of infection hasbeen repeatedly noted with the natural disease(13) and with induced infection (1, 8), but thepresent data emphasize again that clinical infec-tion does not always evoke measurable changesin concentration of serum antibody. The pointof significance here is that while a fairly highcorrelation exists between distinct- clinical re-actions and positive serological responses in agroup of unvaccinated individuals, the correlationbetween the two was very low in a group of vac-cinated subjects. Thus, if antibody responsealone were to serve as index of infection, theevidence would be weighted in favor of the vac-cinated group.

COMPARISONOF SEROLOGICALRESPONSETO

VACCINATION AND INFECTION

Serological reactions were compared in 2groups of subjects one of which received infec-tious virus by way of the respiratory tract andthe other received non-infectious virus by the sub-cutaneous route. The results are shown in Figure5. The group used to illustrate the effect of in-halation of active virus is the unvaccinated con-trol group of 36 individuals. Half of the groupof 41 illustrating the effect of vaccination is com-

542

VACCINATION AGAINST INFLUENZA A

UNVACCINATED CONTROLS

*

-

103

^ 1020.

oi

hir

v

IL

O

x0

x.<9

*

0-0 *-O-

00 o

.* --O * --O0*

a b g

. 0

1~~ ~11 I<32 32 64 I28 25b 512 102420484096

AN TIBODY TIT ER

VACCINATED 4 M OS.BEFORE INFECTION

0---0* -O

* O00

*-O0O* .-O

* O* t° ^O

* **-O0 0

* - * *

0

0

1 1 1 1 I __ 1 __ -__-<32 32 64 128 256 512 1024 2048 4096

ANTIBO DY TI TER

VACCINATED 2 WEEKS

BEFORE INFECTION

102

0

0-0*0

0 0*

* *0

*-o* 0

h

ioz 101

0.hiIL

w 100

3

_0 _03

--O 0-0 -) 99

0. 3

I I I I I I

V AC C I N A TE D WI C E4 MOS. AND 2 WEEKSBEFORE INFECION

LEGENDANTIBODY TITER 0 NO CHANGEANTIBODY TITER * 0

BEFORE AFTER

*-0 0_.0

000 *-O

0 0 00

*-.0

*-- O 00

0

I I

<32 32 64 128 255 512 104 2U48 409O

ANTIBODY TITER

FIG. 4. RELATION OF SEROLoGIcAL AND FEBRILE RESPONSESTO INDUCED INFLUENZATIPF A IN VACCINATED AND CONTROLSUBJECTS

10*

543

e 101

0

LI0. hio1-

100

3

xC 99

103

00.

idM IOLI-

<hi0.3

I-

3

3

2

<32 32 64 128 256 512 1024 2046 4095

ANTIBODY TITER

.4

v -- - - . ..- --.-.- .- - -aAL-

2

la

0 *- * O

T. FRANCIS, JR., J. E. SALK., H. E. PEARSON, AND P. N. BROWN

z

0

~1

Z40 0 00

20

102 0 0

F S12 _ * //

0:~~~~~

F128

0 32 * /

Z <32

<32 32ANTIBODY TITER BEFORE INFECTION

A

I-IL

I.-

c

C

~4

:409

20 *48

1024 * * *

J 512 0 0 0

0E

C2fI **

0

25 * * *00

128. 0

64 0 0

32 0

'<32

<32 32 64 128 256 512 1024

ANTI BODY TITER BEFORE INJECTION

BFIG. 5. A. ANTIBODY TITERS BEFOREAND Arm INHALATION OF AcTIVE INFLUENZA VIRus, TYPE A.

B. CHANGEIN ANTIBODY TITER FOLLOWINGFIRST INJECTION OF VACCINE (TYPE A)

prised of the subjects vaccinated 2 weeks beforeexposure to the virus and the remainder consistsof a group of similarly treated individuals whoparticipated in another study (14). The simi-larities in response to the administration of virus,in either form, by the different routes was quitestriking. The distribution of titers was such thattiters of 128 or less were observed in 90 per centof individuals before vaccination and in 37 per

cent after; in the group receiving active virus byinhalation, 89 per cent had titers of 128 or lessbefore treatment and 36 per cent had this levelof antibody after treatment.

Four-fold or greater increases in antibody titerwere observed in 56 per cent of the vaccinatedgroup and 47 per cent of the group receiving theinhalation; while 2-fold changes in titer occurredin 32 per cent of the vaccinated and 19 per centof the infected groups. Titers were unchangedin 12 per cent of the vaccinated as compared with33 per cent of the group treated with active virus.Although approximately the same proportion ofsubjects reacted with 4-fold or greater increasesin antibody titer following either procedure, pro-

portionately fewer individuals failed to show some

degree of change in titer after vaccination as com-

pared -with infection. The levels to which anti-

body titers were raised were not strikingly dif-ferent. Similar findings have been reported byothers ( 1 ).

DISCUSSION

In earlier studies, Francis (10) had shown theimportance of the quantitative relationships be-tween the immunizing dose of influenza virusand the resultant immunity in ferrets or mice.Certain investigators (15) have reported that,within limits, proportionately higher antibodylevels are obtained when vaccines containing in-creasing amounts of influenza virus are injectedsubcutaneously in humans. While others (1)were able to demonstrate the protective effect ofsubcutaneously injected allantoic fluid vaccine con-

taining influenza virus Type A, against inducedinfection in children, they noted no difference in

the antibody response among groups receivingdifferent amounts of virus. Despite the lack of

direct evidence in human individuals, the possi-bility remained that, with larger amounts of virusin the vaccinating dose, the immunity developedmight be more readily demonstrable. Conse-

quently, in the present studies, a concentratedvaccine was used.

Because the results herein reported were ob-

544

4i4iI

IaII

IaI

c

a

A4

VACCINATION AGAINST INFLUENZA A

tained by experimental infection, their interpre-tation in terms of the natural disease must retaincertain reservations. Nevertheless, by thesemeans, it has been possible to demonstrate greaterresistance to infection with influenza virus, TypeA, in groups of vaccinated subjects than in asimilar group of unvaccinated controls. The rea-son why there was a relatively large number ofpersons who responded with febrile reactions inthe group vaccinated 41/2 months previously can-not be fully explained. It could represent theresults of a failure to develop immunity followingvaccination or it could mean that there had beena decrease of an immunity which originally wasdeveloped in response to inoculation. However,in view of the fact that the groups vaccinated 2weeks before infection were more resistant, theincreased amount of clinical illness in those vac-cinated 4 months before testing suggests that theresult was related to a waning of resistance whichhad been more marked earlier. The question ofduration of immunity following vaccination re-quires further study.

Information regarding the effect of re-inocula-tion in increasing the titer of circulating antibodyand its relation to immmunity is limited to theobservations upon one group of 16 individuals.This has revealed that in a group re-vaccinated 4months after their initial inoculation, 6 of 16, or38 per cent, showed a 2-fold or greater increase intiter, and only 2, or 13 per cent had greater than2-fold increases. Both the latter had titers lessthan 32 prior to the second injection. Failure ofthe second inoculation, given 4 months after thefirst, to provoke a further rise in antibody inmany of the subjects and only a slight rise inothers, may be due in part to combination of anti-gen with antibody, rendering the virus antigeni-cally ineffective. What effect the second inocula-tion may have upon the persistence of the anti-body level is not known but it is clear that theadditional inoculation after 4 months did notfunction as a booster dose.

In the present study of induced influenza, therewas an opportunity to observe the variation inserological response in subjects who developedthe disease. While a fairly high correlation wasfound to exist between distinct clinical responsesand positive serological reaction in a group of un-vaccinated individuals, the correlation between

the two was very low in a group of vaccinatedsubjects. The reason for this dissociation be-tween clinical and serological response is believedto be related, in part at least, to the existence ofrelatively high antibody titers before infection inthe vaccinated groups. It is again evident in thepresent study that experimental influenza canactually be induced without a demonstrable in-crease in serum antibody during convalescence.This observation supports further the notion thata certain number of cases, occurring during out-breaks of the natural disease, without serologicalresponse, still represent infections with influenzavirus.

Serological response in a group of individualspreviously exposed to inhalation of attenuated ac-tive virus has been employed as index of infec-tion (16). For the most part, these subjects hadrelatively high antibody titers after their first ex-posure, and failure of the group as a whole torespond with a second antibody rise was taken tosignify immunity. The results of the presentstudy emphasize the fact that if serological find-ings, rather than clinical observation, are used asindex of infection, the evidence is immediatelyweighted in favor of the vaccinated or recently in-fected group. As a result of this study of theeffectiveness of vaccination against experimentallyinduced influenza A, it has become apparent thatlaboratory criteria of infection alone are not suf-ficient to determine the effectiveness of vaccina-tion against the natural disease.

SUMMARY

The protective effect of a vaccine containinginactivated influenza viruses Types A and Bagainst induced infection with the Type A virushas been studied in man. Distinct clinical reac-tions, evidenced by temperatures of 100° or more,were observed in 50 per cent of the controls, 32per cent of those vaccinated 41/2 months before,14 per cent of those vaccinated 2 weeks before,and 18 per cent of those vaccinated twice. More-over, none in the latter two groups had tempera-tures higher than 100.80, while 25 per cent ofthe controls and 11 per cent of the group vacci-nated 41/2 months before had fever of 1010 orhigher.

Distinct febrile reactions, following exposure to

545

T. FRANCIS, JR., J. E. SALK, H. E. PEARSON, AND P. N. BROWN

virus, occurred in 49 per cent of 53 individualshaving pre-infection antibody titers of 128 or lessand in only 14 per cent of those with titers of256 or more. Moreover, only one of the latterhad as much as 1010, while 11 of the former had1010 or more.

Clinical signs of infection with influenza virusmay occur with or without serological evidence;the latter set of circumstances is most commonlynoted in vaccinated individuals with high anti-body titers. Similarly, serological evidence of in-fection may be obtained with or without obviousclinical signs.

It was noted that a second inoculation of vac-cine, 4 months after the first, did not elicit afurther sharp rise in antibodies.

The question of duration of immunity followingvaccination has been discussed.

BIBLIOGRAPHY

1. Henle, W., Henle, G., and Stokes, J., Jr., Demonstra-tion of the efficacy of vaccination against influenzatype A by experimental infection of human beings.J. Immunol., 1943, 46, 163.

2. Francis, T., Jr., Salk, J. E, Pearson, H. E., andBrown, P. N., Protective effect of vaccinationagainst induced influenza A. Proc. Soc. Exper.Biol. and Med., 1944, 55, 104.

3. Francis, T., Jr., Transmission of influenza by a filter-able virus. Science, 1934, 80, 457.

4. Francis, T., Jr., A new type of virus from epidemicinfluenza. Science, 1940, 92, 405.

5. Francis, T., Jr., and Salk, J. E., A simplified proce-dure for the concentration and purification of influ-enza virus. Science, 1942, 96, 499.

6. Hirst, G. K., The quantitative determination of influ-enza virus and antibodies by means of red cell ag-glutination. J. Exper. Med., 1942, 75, 49.

7. Salk,- J. E., A simplified procedure for titratinghemagglutinating capacity of influenza-virus andthe corresponding antibody. J. Immunol., 1944,49, 87.

8. Francis, T., Jr., Pearson, H. E., Salk, J. E., andBrown, P. N., Immunity in human subjects arti-ficially infected with influenza virus, Type B. Am.J. Pub. Health, 1944, 34, 317.

9. Taylor, R. M., Experimental infection with influenzaA virus in mice. The increase in intrapulmonaryvirus after inoculation and the influence of variousfactors thereon. J. Exper. Med., 1941, 73, 43.

10. Francis, T., Jr., Quantitative relationships betweenthe immunizing dose of epidemic influenza virusand the resultant immunity. J. Exper. Med., 1939,69, 283.

11. Studies to be published.12. Horsfall, F. L., Jr., Lennette, E. H., Rickard, E. R.,

and Hirst, G. K., Studies on the efficacy of a com-plex vaccine against influenza A. Pub. HealthRep., 1941, 56, 1863.

13. Francis, T., Jr., Magill, T. P., Rickard, E. R., andBeck, M. D., Etiological and serological studies inepidemic influenza. Am. J. Pub. Health, 1937, 27,1141.

14. Salk, J. E., Pearson, H. E., Brown, P. N., and Fran-cis, T., Jr., Protective effect of vaccination againstinduced influenza B. Proc. Soc. Exper. Biol. andMed., 1944, 55, 106.

15. Hirst, G. K., Rickard, E. R., Whitman, L., and Hors-fall, F. L., Jr., Antibody response of human beingsfollowing vaccination with influenza viruses. J.Exper. Med., 1942, 75, 495.

16. Bull, D. R., and Burnet, F. M., Experimental immu-nization of volunteers against influenza virus B.M. J. Australia, 1943, 1, 389.

546